Sampling Plan: The Complete Guide to Pharmaceutical Acceptance Sampling
A sampling plan is a documented quality control procedure that specifies how many units to inspect from a lot, what acceptance criteria to use, and how to decide whether to accept or reject the batch. Proper sampling plans balance statistical confidence with inspection efficiency, ensuring products meet specifications while optimizing resources.
A sampling plan is a documented procedure that specifies the sample size, acceptance criteria, and decision rules used to determine whether a lot or batch meets quality specifications based on the inspection of a representative sample. In pharmaceutical manufacturing, sampling plans form the foundation of quality control programs, ensuring that released products consistently meet safety and efficacy standards while optimizing inspection resources.
Every quality professional in pharma and biotech faces the same fundamental challenge: you cannot inspect every unit in a production lot, yet you must be confident that the entire lot meets specifications. Statistical sampling provides the scientific framework for making accept/reject decisions with quantifiable confidence levels - but only when the sampling plan is correctly designed and implemented.
The consequences of inadequate sampling plans extend far beyond rejected lots. Poorly designed sampling plans can either release defective product to patients (consumer risk) or unnecessarily reject good lots (producer risk). In regulated industries, sampling plan deficiencies are frequently cited in FDA 483 observations and warning letters, making proper sampling plan design both a quality imperative and a compliance requirement.
In this guide, you will learn:
- What sampling plans are and why they matter for pharmaceutical quality
- The key standards governing acceptance sampling, including ANSI/ASQ Z1.4
- How to determine appropriate AQL levels and sample sizes
- The difference between attribute and variable sampling plans
- How to select inspection levels for different quality characteristics
- FDA and GMP requirements for pharmaceutical sampling plans
What Is a Sampling Plan?
Sampling Plan - A formal, documented quality control procedure that specifies the sample size (number of units to inspect), acceptance number (maximum defects allowed), and decision rules for accepting or rejecting a lot based on inspection results. Sampling plans apply statistical methods to make confident accept/reject decisions without inspecting every unit.
A sampling plan is a formal, documented scheme that defines the number of units to inspect from a lot (sample size), the maximum number of defects or defective units permitted (acceptance number), and the criteria for accepting or rejecting the lot based on inspection results. Sampling plans translate statistical theory into practical quality control procedures.
Key components of a sampling plan:
| Component | Definition | Example |
|---|---|---|
| Lot size (N) | Total units in the batch being evaluated | 10,000 tablets |
| Sample size (n) | Number of units selected for inspection | 200 tablets |
| Acceptance number (Ac) | Maximum defects to accept the lot | 5 defects |
| Rejection number (Re) | Minimum defects to reject the lot | 6 defects |
| AQL | Acceptable Quality Level - maximum tolerable defect rate | 1.0% |
“Key Principle: A sampling plan does not guarantee zero defects in a lot. Instead, it provides statistical confidence that the true defect rate in the lot is at or below the specified AQL with a defined probability.
Sampling plans operate on the principle that a properly selected random sample reflects the quality characteristics of the entire lot. By applying statistical methods, quality professionals can make informed accept/reject decisions without the prohibitive cost and time of 100% inspection.
Types of Sampling Plans
Sampling plans are categorized by how samples are taken and what characteristics are measured.
| Plan Type | Description | When to Use |
|---|---|---|
| Single sampling | One sample, one decision | Most common, simple implementation |
| Double sampling | Initial sample, conditional second sample | When cost of sampling is high |
| Multiple sampling | Series of smaller samples until decision | Complex scenarios, large lots |
| Sequential sampling | One unit at a time until decision | Destructive testing, high-value items |
By characteristic measured:
| Classification | What Is Measured | Example Tests |
|---|---|---|
| Attribute sampling | Pass/fail, defective/non-defective | Visual defects, go/no-go gauges |
| Variable sampling | Actual measured values | Tablet weight, dissolution, assay |
Understanding these classifications is essential because different sampling standards apply to attribute versus variable plans, and the sample size requirements differ significantly.
Acceptance Sampling Standards: ANSI/ASQ Z1.4 and Z1.9
The two primary standards governing acceptance sampling in pharmaceutical and other industries are ANSI/ASQ Z1.4 for attribute sampling and ANSI/ASQ Z1.9 for variable sampling. These standards provide scientifically validated sampling tables and procedures.
ANSI/ASQ Z1.4: Attribute Sampling Standard
ANSI/ASQ Z1.4 (identical to ISO 2859-1 and originally MIL-STD-105E) is the most widely used acceptance sampling standard for attribute data. It provides sampling plans indexed by Acceptable Quality Level (AQL) for inspection by attributes.
ANSI/ASQ Z1.4 key features:
| Feature | Description |
|---|---|
| AQL-indexed | Plans organized by acceptable quality level (0.010% to 10%) |
| Three inspection levels | General levels I, II, III for different discrimination |
| Four special levels | S-1, S-2, S-3, S-4 for small sample requirements |
| Switching rules | Normal, tightened, reduced inspection based on quality history |
| Lot size ranges | 2 to 500,000+ units with corresponding sample sizes |
When to use ANSI/ASQ Z1.4:
- Inspecting for defects that are classified as defective/non-defective
- Counting the number of defects per unit
- Visual inspection for appearance defects
- Functional testing with pass/fail criteria
- Container/closure integrity testing
ANSI/ASQ Z1.9: Variable Sampling Standard
ANSI/ASQ Z1.9 (similar to ISO 3951) provides sampling plans for inspection by variables - when actual measurements are taken and evaluated against specifications.
ANSI/ASQ Z1.9 key features:
| Feature | Description |
|---|---|
| Measurement-based | Uses actual measured values, not just pass/fail |
| Smaller samples | Typically requires fewer samples than attribute plans |
| Assumes normality | Data must follow normal distribution |
| Two methods | Standard deviation method and range method |
| AQL-indexed | Same AQL framework as Z1.4 |
When to use ANSI/ASQ Z1.9:
- Quantitative test results (assay, weight, dimensions)
- When smaller sample sizes are economically important
- When data is normally distributed
- For characteristics with numerical specifications
Comparison: ANSI/ASQ Z1.4 vs Z1.9
| Aspect | Z1.4 (Attributes) | Z1.9 (Variables) |
|---|---|---|
| Data type | Pass/fail, counts | Measurements |
| Sample size | Larger | Smaller (30-50% reduction typical) |
| Information extracted | Defect count only | Mean, variability, process capability |
| Distribution assumption | None | Normal distribution required |
| Calculation complexity | Simple counting | Statistical calculations required |
| Cost per unit | Lower testing cost | Higher measurement cost |
| Application | Visual defects, functional tests | Lab tests, dimensional checks |
When choosing between Z1.4 and Z1.9, ask yourself: "Is the measurement already being taken?" If your lab is already measuring tablet weight, assay, or dissolution for every sample, use Z1.9 variable sampling to reduce sample sizes by 30-50%. This directly reduces inspection costs while maintaining equivalent confidence.
“Industry Practice: Many pharmaceutical quality control programs use variable sampling (Z1.9) for laboratory testing where measurements are already taken, and attribute sampling (Z1.4) for visual inspection and packaging verification.
AQL Sampling: Understanding Acceptable Quality Levels
At the AQL quality level, approximately 95% of lots will be accepted and 5% will be rejected. This 5% rejection rate is the producer's risk-a natural consequence of statistical sampling, not a sign of poor quality. Understanding this distinction prevents premature process changes based on normal sampling variation.
The Acceptable Quality Level (AQL) is the maximum defect rate considered acceptable as a process average when a continuing series of lots is submitted for acceptance sampling. AQL is the central concept in acceptance sampling and directly determines sample size requirements.
AQL Definition and Interpretation
| Term | Definition | Interpretation |
|---|---|---|
| AQL | Acceptable Quality Level | The quality level considered satisfactory for the process |
| Producer's Risk (alpha) | Probability of rejecting a good lot | Typically 5% at AQL level |
| Consumer's Risk (beta) | Probability of accepting a bad lot | Typically 10% at specified LQL |
| LQL/LTPD | Lot Quality Level/Lot Tolerance Percent Defective | Quality level with 10% acceptance probability |
Critical distinction: AQL does not mean lots at the AQL level are always accepted. At the AQL quality level, approximately 95% of lots will be accepted and 5% rejected. This 5% rejection probability is the producer's risk.
Selecting Appropriate AQL Levels
AQL selection depends on the defect classification and its impact on product quality, patient safety, and regulatory compliance.
Typical AQL levels by defect classification:
| Defect Class | Description | Typical AQL | Examples |
|---|---|---|---|
| Critical | Could cause harm or safety issues | 0% (100% inspection) | Contamination, wrong drug |
| Major | Affects efficacy or function | 0.10% - 0.65% | Potency failures, dissolution issues |
| Minor | Cosmetic or administrative | 1.0% - 4.0% | Labeling appearance, minor visual defects |
AQL selection considerations for pharmaceuticals:
- Patient safety impact - Critical defects affecting patient safety typically require 100% inspection or very stringent AQL (0.010% - 0.065%)
- Regulatory requirements - FDA and other agencies may specify minimum sampling requirements for certain tests
- Process capability - AQL should be achievable by the manufacturing process under normal conditions
- Historical quality data - Past performance informs realistic AQL targets
- Customer requirements - Contract manufacturers may need to meet customer-specified AQL levels
Don't set AQL tighter than your process can consistently achieve. If your process naturally produces a 0.5% defect rate, setting AQL at 0.10% will trigger frequent rejections and frustration. Instead, set AQL based on three factors: (1) safety impact (how much does this defect matter?), (2) regulatory expectations (what does FDA expect?), and (3) actual process capability (what can we realistically achieve?).
AQL Sample Size Relationship
The relationship between AQL, lot size, and sample size is defined in sampling tables. Lower AQL values require larger sample sizes.
Sample size comparison by AQL (General Inspection Level II):
| Lot Size | AQL 0.10% | AQL 0.65% | AQL 1.0% | AQL 2.5% |
|---|---|---|---|---|
| 501-1,200 | 200 | 80 | 80 | 50 |
| 1,201-3,200 | 315 | 125 | 125 | 80 |
| 3,201-10,000 | 500 | 200 | 200 | 125 |
| 10,001-35,000 | 800 | 315 | 315 | 200 |
| 35,001-150,000 | 1,250 | 500 | 500 | 315 |
This table illustrates a fundamental principle: tighter quality requirements (lower AQL) demand more extensive sampling to achieve the same confidence in lot acceptance decisions.
Statistical Sampling: Sample Size Determination Methods
Statistical sampling applies probability theory to determine sample sizes and acceptance criteria that achieve desired confidence levels in lot acceptance decisions. Understanding sample size determination is essential for designing effective sampling plans.
Using variable sampling instead of attribute sampling reduces sample sizes by 30-50% while maintaining equivalent statistical confidence. For a 10,000-unit lot with AQL 1.0%, attribute sampling requires 200 units; variable sampling requires only 50 units. This directly translates to cost savings in inspection time and resources.
Factors Affecting Sample Size
| Factor | Effect on Sample Size | Explanation |
|---|---|---|
| AQL (tighter) | Increases sample size | More samples needed to detect smaller defect rates |
| Lot size (larger) | Increases sample size | But relationship is not linear |
| Inspection level (higher) | Increases sample size | Greater discrimination requires more data |
| Producer's risk (lower) | Increases sample size | More confidence in accepting good lots |
| Consumer's risk (lower) | Increases sample size | More protection against accepting bad lots |
Sample Size Determination Process
For ANSI/ASQ Z1.4 (Attribute Sampling):
- Determine lot size (N) - Count total units in the lot
- Select inspection level - Usually General Level II unless specified otherwise
- Find sample size code letter - Use lot size and inspection level in Table I
- Identify AQL - Based on defect classification and requirements
- Look up sample size (n) and acceptance number (Ac) - Use Table II-A for single sampling
Example calculation:
| Input | Value |
|---|---|
| Lot size | 5,000 tablets |
| Inspection level | General Level II |
| AQL | 1.0% |
| Sample size code | L |
| Sample size (n) | 200 |
| Acceptance number (Ac) | 5 |
| Rejection number (Re) | 6 |
Decision rule: Inspect 200 tablets randomly selected from the lot. Accept the lot if 5 or fewer defective tablets are found. Reject if 6 or more defective tablets are found.
Operating Characteristic (OC) Curves
The Operating Characteristic curve shows the probability of lot acceptance at various quality levels. Understanding OC curves is essential for evaluating sampling plan effectiveness.
OC curve interpretation:
| Quality Level | Acceptance Probability | Meaning |
|---|---|---|
| At AQL | ~95% | Good lots usually accepted |
| At IQL (Indifference) | ~50% | Equal chance accept/reject |
| At LQL/LTPD | ~10% | Poor lots usually rejected |
Key OC curve metrics:
- Producer's risk (Type I error) - Probability of rejecting a lot at AQL quality (~5%)
- Consumer's risk (Type II error) - Probability of accepting a lot at LQL quality (~10%)
- Curve steepness - Steeper curves better discriminate between good and bad lots
“Practical Insight: Larger sample sizes produce steeper OC curves, meaning better discrimination between acceptable and unacceptable quality levels. However, this comes at increased inspection cost.
Sampling Plan Pharmaceutical: FDA and GMP Requirements
Sampling plan pharmaceutical requirements are defined by FDA regulations, GMP guidelines, and pharmacopeial standards. Pharmaceutical sampling plans must satisfy both statistical validity and regulatory compliance.
FDA Sampling Requirements
FDA 21 CFR Part 211 (Current Good Manufacturing Practice) establishes sampling requirements for pharmaceutical manufacturing:
21 CFR 211.84 - Testing and approval/rejection of components, drug product containers, and closures:
| Requirement | Description |
|---|---|
| Representative samples | Samples must represent the lot being tested |
| Written procedures | Sampling procedures must be documented |
| Sample quantity | Sufficient for all required tests |
| Container identification | Each container sampled must be identified |
| Statistical criteria | "Representative and adequate" sampling |
21 CFR 211.165 - Testing and release for distribution:
| Requirement | Description |
|---|---|
| Batch testing | Each batch must be tested for conformance |
| Sampling procedures | Written procedures describing sampling methods |
| Statistical confidence | "Adequate" confidence in batch quality |
| Special considerations | Sterile products require specific sampling |
GMP Sampling Plan Requirements
GMP guidelines from FDA, EU, and WHO provide additional sampling guidance:
Key GMP sampling principles:
- Risk-based approach - Sampling intensity based on quality risk assessment
- Statistical validity - Plans based on recognized statistical methods
- Representative sampling - All portions of the lot have equal chance of selection
- Documentation - Complete records of sampling procedures and results
- Trend analysis - Use sampling data for ongoing quality monitoring
Pharmaceutical-Specific AQL Guidelines
| Test Category | Typical AQL Range | Rationale |
|---|---|---|
| Identity testing | 0% (100% testing) | Patient safety critical |
| Potency/Assay | 0.10% - 0.65% | Efficacy-related |
| Dissolution | 0.65% - 1.0% | Bioavailability impact |
| Content uniformity | 0.65% | Dose consistency |
| Particulate matter | 0.10% - 0.65% | Safety (parenteral products) |
| Container closure | 1.0% - 2.5% | Integrity protection |
| Labeling/packaging | 1.0% - 4.0% | Administrative, patient safety |
| Visual defects | 2.5% - 6.5% | Appearance, minor impact |
USP Sampling Guidance
USP General Chapter <1790> provides guidance on visual inspection of parenterals, with specific sampling recommendations:
| Inspection Type | Sample Size | Acceptance Criteria |
|---|---|---|
| 100% inspection | All units | Zero defects (critical) |
| Statistical sampling | Per AQL tables | Based on defect class |
| Skip-lot testing | Based on history | For qualified processes |
Attribute vs Variable Sampling Plans: Choosing the Right Approach
Understanding when to use attribute versus variable sampling plans is critical for efficient and effective quality control.
Attribute Sampling Plans
Attribute sampling classifies each unit as either conforming or non-conforming (good or defective) without measuring actual values.
Attribute sampling characteristics:
| Aspect | Description |
|---|---|
| Measurement | Pass/fail, acceptable/defective |
| Sample size | Larger than variable plans |
| Calculation | Count defectives, compare to acceptance number |
| Skills required | Trained inspectors, clear defect definitions |
| Cost | Lower per-unit inspection cost |
Best applications for attribute sampling:
- Visual inspection for appearance defects
- Go/no-go functional testing
- Presence/absence verification (label, seal, components)
- Container closure integrity (deterministic testing)
- Foreign particulate inspection
Attribute sampling example:
| Parameter | Value |
|---|---|
| Characteristic | Label legibility |
| Lot size | 50,000 units |
| AQL | 2.5% |
| Sample size | 200 |
| Acceptance number | 10 |
| Test method | Visual inspection |
| Decision | Count illegible labels, accept if 10 or fewer |
Variable Sampling Plans
Variable sampling measures actual values and compares them against specifications using statistical calculations.
Variable sampling characteristics:
| Aspect | Description |
|---|---|
| Measurement | Actual values recorded |
| Sample size | Smaller than attribute plans (typically 30-50% less) |
| Calculation | Mean, standard deviation, quality indices |
| Skills required | Statistical calculations, measurement capability |
| Cost | Higher per-unit measurement cost |
Best applications for variable sampling:
- Weight verification
- Dimensional measurements
- Assay/potency testing
- Dissolution testing
- pH measurements
- Any quantitative laboratory testing
Variable sampling example:
| Parameter | Value |
|---|---|
| Characteristic | Tablet weight |
| Specification | 500 mg plus/minus 5% (475-525 mg) |
| Lot size | 50,000 tablets |
| AQL | 1.0% |
| Sample size | 50 |
| Method | Calculate mean and standard deviation |
| Acceptance criteria | Quality index exceeds critical value |
Comparison Summary: Attribute vs Variable
| Factor | Attribute Sampling | Variable Sampling |
|---|---|---|
| Data collected | Pass/fail only | Actual measurements |
| Sample size | Larger | Smaller (30-50% reduction) |
| Information gained | Lot conformance | Conformance + process data |
| Distribution assumption | None | Normal distribution |
| Calculation complexity | Simple counting | Statistical calculations |
| Process improvement data | Limited | Rich data for trending |
| Inspector training | Defect recognition | Measurement technique |
| Equipment needs | Visual aids, gauges | Calibrated instruments |
“Best Practice: When measurement data is already being collected (such as laboratory testing), use variable sampling plans to reduce sample sizes. Reserve attribute sampling for truly categorical characteristics like visual defects.
Inspection Levels and Sample Size Code Letters
Inspection levels provide flexibility to adjust sample size based on the discrimination required and the relative cost of inspection versus the cost of passing defective product.
General Inspection Levels
ANSI/ASQ Z1.4 defines three general inspection levels:
| Level | Description | Sample Size | When to Use |
|---|---|---|---|
| Level I | Reduced discrimination | Smallest | Low cost of defectives, expensive inspection |
| Level II | Normal discrimination | Standard | Default level for most applications |
| Level III | Tighter discrimination | Largest | High cost of defectives, critical characteristics |
Sample size comparison by inspection level (Lot size: 10,000):
| Inspection Level | Code Letter | Sample Size |
|---|---|---|
| Level I | H | 50 |
| Level II | L | 200 |
| Level III | N | 500 |
Special Inspection Levels
Four special inspection levels (S-1, S-2, S-3, S-4) provide smaller sample sizes for situations where:
- Inspection is destructive
- Testing costs are very high
- Small sample sizes are acceptable due to process history
| Special Level | Sample Size Range | Typical Use |
|---|---|---|
| S-1 | Very small | Destructive testing, extremely high cost |
| S-2 | Small | Destructive testing |
| S-3 | Moderate | Expensive testing |
| S-4 | Approaches Level I | Costly but important testing |
Selecting the Appropriate Inspection Level
| Factor | Favors Lower Level | Favors Higher Level |
|---|---|---|
| Inspection cost | High cost | Low cost |
| Defective cost | Low impact | High impact |
| Process history | Stable, capable | Variable, new |
| Regulatory requirement | Not specified | Mandated |
| Risk tolerance | Higher acceptable | Lower acceptable |
Switching Rules: Normal, Tightened, and Reduced Inspection
ANSI/ASQ Z1.4 includes switching rules that adjust inspection intensity based on quality history. These rules reward good quality with reduced inspection and penalize poor quality with tightened requirements.
Switching Rules Summary
| Inspection State | Trigger to Enter | Sample Size | Acceptance Criteria |
|---|---|---|---|
| Normal | Starting point | Standard | Standard |
| Tightened | 2 of 5 lots rejected | Standard | Stricter (fewer defects allowed) |
| Reduced | 10 consecutive lots accepted + production steady | Smaller | Standard or relaxed |
Normal to Tightened Switching
Switch from normal to tightened inspection when:
- 2 out of 5 consecutive lots are rejected on original inspection
Effects of tightened inspection:
- Same sample size as normal
- Lower acceptance numbers (stricter criteria)
- Approximately 1.6x more likely to reject lots at same quality
Tightened to Normal Switching
Switch from tightened back to normal when:
- 5 consecutive lots accepted under tightened inspection
Normal to Reduced Switching
Switch from normal to reduced inspection when ALL conditions met:
- 10 consecutive lots accepted under normal inspection
- Total defects in those lots is at or below limit number
- Production is steady (no process changes)
- Approved by responsible authority
Effects of reduced inspection:
- Smaller sample sizes (typically 40% of normal)
- Acceptance numbers may be fractional or special criteria
- Provides cost savings reward for consistent quality
Reduced to Normal Switching
Switch from reduced back to normal when ANY occurs:
- Lot rejected under reduced inspection
- Production becomes irregular
- Other conditions warrant
Discontinuation of Inspection
If 5 consecutive lots rejected under tightened inspection:
- Discontinue acceptance sampling
- Investigate and correct the process
- Resume only when improvement demonstrated
“Implementation Note: Switching rules require careful tracking of lot history. Many pharmaceutical companies maintain electronic batch records that automatically track switching status and alert quality personnel when switches are required.
Implementing Sampling Plans in Pharmaceutical Quality Systems
Successful implementation of sampling plans requires integration with broader quality management systems and clear documentation.
Sampling Plan Documentation Requirements
| Document Element | Content |
|---|---|
| Scope | Which products, processes, tests covered |
| Sampling standard | ANSI/ASQ Z1.4, Z1.9, or other reference |
| AQL levels | By defect class and characteristic |
| Inspection levels | General/special level selection rationale |
| Sample size tables | Or reference to standard tables |
| Acceptance criteria | Clear accept/reject decision rules |
| Switching rules | If applicable, conditions for each switch |
| Sampling procedures | How to select representative samples |
| Record requirements | What data to capture and retain |
Sample Selection Methods
Proper random sampling is essential for statistical validity:
| Method | Description | When to Use |
|---|---|---|
| Simple random | Each unit equally likely | Homogeneous lots |
| Stratified random | Random within defined subgroups | Known variation sources |
| Systematic | Every nth unit after random start | Production line sampling |
| Cluster | Random selection of groups | Container-based lots |
Random number generation: Use validated random number generators or published random number tables. Avoid subjective or convenience sampling.
Integration with Batch Release
Sampling plans connect to the broader batch release process:
- Batch record review - Verify sampling performed per approved plan
- Results evaluation - Compare results to acceptance criteria
- OOS investigation - If failures, investigate per OOS procedure
- Lot disposition - Release, reject, or hold based on cumulative data
- Trending - Accumulate data for process monitoring
Common Implementation Pitfalls
| Pitfall | Impact | Prevention |
|---|---|---|
| Non-random sampling | Invalid statistical basis | Use validated random selection |
| Wrong sample size | Over/under sampling | Verify against lot size and AQL |
| Ignoring switching rules | Non-compliance with standard | Track lot history systematically |
| Unclear defect definitions | Inconsistent classification | Train inspectors, use visual standards |
| Inadequate documentation | Audit findings | Complete records for each lot |
Key Takeaways
A sampling plan is a documented procedure that specifies how samples are selected from a lot and how the lot acceptance decision is made based on inspection results. It defines the sample size (number of units to inspect), acceptance number (maximum defects allowed), and rejection number (minimum defects triggering rejection). Sampling plans are based on statistical methods that provide quantifiable confidence levels for accept/reject decisions.
Key Takeaways
- A sampling plan is a documented quality control procedure that specifies sample size, acceptance criteria, and decision rules for lot acceptance based on statistical principles. Proper sampling plans balance inspection costs against quality risks.
- ANSI/ASQ Z1.4 and Z1.9 are the primary acceptance sampling standards for attribute and variable data respectively. Z1.4 is used for pass/fail inspections; Z1.9 for measured values with smaller sample sizes.
- AQL (Acceptable Quality Level) determines sampling stringency. Lower AQL values require larger sample sizes but provide greater confidence. AQL selection should reflect defect criticality and patient safety impact.
- Variable sampling plans require fewer samples than attribute plans (typically 30-50% reduction) but require normal distribution assumption and statistical calculations. Use variable plans when measurement data is already collected.
- Inspection levels provide sampling flexibility. Level II is standard; Level I for expensive inspection/low-risk defects; Level III for critical characteristics. Special levels S-1 through S-4 accommodate destructive or costly testing.
- Switching rules reward quality consistency. Normal inspection switches to tightened after rejections, to reduced after consistent acceptance. Proper tracking ensures compliance with switching protocols.
- ---
Next Steps
Effective sampling plans are not just statistical exercises - they are critical quality control tools that protect patient safety while optimizing inspection resources. The key is selecting the right AQL levels, inspection levels, and sampling methods for your specific products and quality characteristics.
Build confidence in your batch release decisions. Assyro's AI-powered quality intelligence platform helps regulatory and quality teams implement statistically sound sampling plans integrated with comprehensive batch documentation. Our technology ensures sampling plan compliance while providing the data transparency needed for regulatory inspections and audits.
See How Assyro Supports Quality Control Documentation - Request a Demo
Sources
Sources
- ANSI/ASQ Z1.4-2008: Sampling Procedures and Tables for Inspection by Attributes
- ANSI/ASQ Z1.9-2008: Sampling Procedures and Tables for Inspection by Variables
- FDA 21 CFR Part 211: Current Good Manufacturing Practice for Finished Pharmaceuticals
- USP General Chapter <1790> Visual Inspection of Injections
- ISO 2859-1:1999 Sampling procedures for inspection by attributes
- FDA Guidance for Industry: Quality Systems Approach to Pharmaceutical CGMP Regulations